Treatment Planning Comparison of Volumetric Modulated Arc Therapy With Trilogy and Halcyon for Bilateral Breast Cancer

Background: To evaluate the dosimetry of Halcyon in treatment of bilateral breast cancer with volumetric modulated arc therapy. Methods: On CT images of 10 patients with bilateral breast cancer, four Halcyon plans with different setup elds were generated and dosimetric comparisons were conducted among the four plans to select an optimal setup eld mode. The four setup-eld plans were referred to as CBCT-H, CBCT-L, MV-H, MV-L. Whole and partial arc plans on Trilogy and Halcyon referred to as T-4arc, T-8arc, H-4arc and H-8arc were designed. The dosimetric differences between whole and partial arc plans in the same accelerator were compared to understand the most suitable eld setting mode. The better Halcyon plan was selected to the further dosimetric comparison of the plan quality and delivery eciency between Trilogy and Halcyon. Results: CBCT-H plans increased D mean , D 2 and V 107 of planning target volume (PTV) and V 5 and D mean of the heart, left ventricle (LV) and lungs compared to other plans. No signicantly dosimetric differences were observed in PTV and organs at risk (OARs) among CBCT-L, MV-H and MV-L. The mean dose and low dose volume of heart, lungs and liver were signicantly decreased in T-8arc plans. In terms of V 5 , V 20 , V 30 , V 40 and D mean of the heart, V 20 , V 30 , V 40 and D mean of the LV, V 30 , V 40 , D max and D mean of the left anterior descending artery (LAD), V 5 and V 40 of lungs, H-8arc was signicantly higher than H-4arc (p<0.05). Compared to Trilogy’s plans, Halcyon’s plans reduced the high-dose volume of the heart and LV, but increased the mean dose of the heart. For the dose of the LAD and the V 20 , V 30 of lungs, there was no statistical difference between the two accelerators. Compared with Trilogy, plans on Halcyon signicantly increased the skin dose, but also signicantly reduced the delivery time. cone left LAD: left

and daily image guided radiotherapy (IGRT). Comparable dose distribution and excellent delivery e ciency have been shown on Halcyon in radiotherapy for head and neck, brain, unilateral breast, and cervical cancers [3][4][5][6][7]. Trilogy is a conventional C-arm LINAC with jaw and Millennium MLC.
It's proved volumetric modulated arc therapy (VMAT) can improve the quality of the plan and the delivery e ciency in bilateral breast cancer [8][9][10]. Previous study found that the partial-arc VMAT plans signi cantly reduced the cardiopulmonary dose compared to the whole-arc VMAT plans in unilateral breast radiotherapy [11]. Similarly, partial-arc VMAT plans were feasible in bilateral breast radiotherapy [12]. These studies were based on conventional C-arm LINAC with jaw. To our knowledge, the application of Halcyon in the VMAT plans for bilateral breast cancer have not been reported. It is unknown whether the partial-arc VMAT plan is applicable on Halcyon. In Halcyon 1.0, high-quality/low-dose megavoltage cone beam computed tomography (MV CBCT) or orthogonal MV radiograph pair could be selected for image guided, and the four setup-elds delivered different monitor units (MUs). It is important to choose an optimal setup eld.
The aim of the study was to select an optimal setup eld and the suitable arc mode of VMAT plans for Halcyon in bilateral breast radiotherapy, and analyze the plan quality and delivery e ciency by comparing the dosimetric differences between the Halcyon and Trilogy to guide the clinical application in bilateral breast radiotherapy.

Patient Selection and Volume Delineation
From September 2006 to December 2018, CT image datasets of 10 patients diagnosed with bilateral breast cancer and received bilateral breast radiotherapy at Shandong Cancer Hospital were selected. The clinical target volume (CTV) includes all bilateral breast tissue, excluding local lymph node region. The planning target volume (PTV) was generated by expanding a 5-mm margin from the CTV and was shrinked to 5 mm below the skin on the skin side. The organs at risk (OARs) include the total lung, heart, left ventricle (LV), left anterior descending artery (LAD) and liver. The skin is de ned as the 3mm region below the body outside of the PTV. The normal structures were de ned as the body minus the PTV (B-P).

Treatment Planning
In Halcyon version 1.0, all imaging setup elds are taken using digital megavoltage imaging (DMI) panels. When designing a Halcyon plan, four different setup elds could be selected. For the 10 patients, we designed four VMAT plans with different setup elds on Halcyon: high-quality MV cone beam CT (the gantry rotates clockwise from 260° to 100° and delivering 10 monitor units, simply called CBCT-H); lowdose MV CBCT (delivering 5 MUs in a clockwise gantry rotation from 260° to 100°, simply called CBCT-L); high-quality orthogonal MV radiograph pair (images acquired with 0° and 90° and delivering 2 MUs for each eld, simply called MV-H) and low-dose orthogonal MV radiograph pair (images acquired with 0°a nd 90° and delivering 1 MU for each eld, simply called MV-L). For the four plans, two anticlockwise 160° to 200°and two clockwise 200° to 160° rotation arcs were used. First, the dosimetric differences between the four plans were compared to nd the optimal setup mode.
Then, four VMAT plans were designed for the 10 patients. On Trilogy, whole and partial arc plans were generated. The whole-arc plan consisted of two anticlockwise 160° to 200°and two clockwise 200° to 160° rotation arcs. The partial-arc plan consisted of total 8 partial arcs. For unilateral breast, four 100°a round arcs were generated. In the two plans, the medial x-jaw was set to the minimum site (−2 cm) to minimize the irradiated volume of the lungs and heart. 6 MV X ray was used and the dose rate was set to 600 MU/min. The whole and partial arc plans designed on the Trilogy were referred to as T-4arc and T-8arc respectively. Two whole and partial arc plans, referred to as H-4arc and H-8arc, were designed on Halcyon with the same arc angle as T-4arc and T-8arc mentioned above. In Halcyon plans, 6 MV FFF X ray was used and the maximum dose rate was 800 MU/min. Low-quality MVCBCT was selected for image guided.
The prescription dose was 50 Gy in 2-Gy fractions. All plans were designed with the Eclipse version 15.5 treatment planning system (Varian Medical Systems, Palo Alto, CA, USA) using analytic anisotropic algorithm (AAA). The normalization of dose was 95% volume of PTV achieved 100% prescription dose. All plans used the same optimization parameter setting, with the goal of minimizing the dose to the lungs, heart and LAD while ensuring the PTV dose coverage. No dose constraint was applied to the skin and LV during the optimization.

Dosimetric Evaluation
The dose statistics of the plans were based on dose-volume histogram (DVH) analysis. For PTV, the dose of 2% and 98% volume (D 2 , D 98 ), the volume receiving 107% and 110% of the prescribed dose (V 107 and V 110 and the mean dose (D mean ) were analyzed. The conformity index (CI) and the homogeneity index (HI) of the PTV were calculated according to the following formula: TV PV represents the volume of the PTV wrapped by the prescription dose, the TV represents the volume of the PTV, and the PV represents the total volume wrapped by the prescription dose. Larger CI values indicate the better conformity of target [13]. D 2 and D 98 represents the irradiated dose of 2% and 98% volume of the PTV respectively. D p represents the prescribed dose. Lower HI values indicate the better uniformity of target [14].
For OARs, the V X and mean doses were analyzed. V X represents the irradiated volume of X Gy dose.
The number of MUs was analyzed for all plans. The delivery time of T-4arc, T-8arc, H-4arc and H-8arc was recorded. The delivery time was recorded from the rst eld beam on to the last eld beam off, excluding the positioning time.

Statistical Analysis
All data were statistically analyzed using the Statistical Package for Social Sciences v20.0 software (SPSS Inc., Chicago, IL, USA). First, A one-way analysis of variance (ANOVA) using Bonferroni's multiple comparisons test were applied to compare the four different setup-eld plans on Halcyon. Second, whole and partial arc plans on same LINAC were compared to determine the most suitable eld mode for the machine. The Mann-Whitney U rank and test was used. The better Halcyon plan was selected to the further comparison. Third, a statistical comparison of the better Halcyon plan and the Trilogy plan was implemented to analyze the dosimetric differences between the two machines using Mann-Whitney U rank and test. The differences were considered statistically signi cant when p < 0.05.

Results
Dose Comparisons for Different Setup-eld Plans Table 1  were observed in other dosimetric parameters among the four plans. Figure 1 shows the dose distributions of the four different setup-eld plans on Halcyon.   D 98 and CI in H-4arc plans were higher than those in H-8arc plans, and the differences were statistically signi cant. There was no statistical difference in D mean of PTV between the two plans on Halcyon.   Figure 2 shows the mean dose-volume histograms for the four plans.
According to the results above, T-8arc and H-4arc plans showed better dosimetry on the respective LINAC. Because of the poor dosimetry, H-8arc was not used for further statistical comparison. Next, H-4arc was statistically analyzed with T-4arc and T-8arc respectively to re ect the dosimetric differences between the two LINACs.  Table 2.
For the doses of heart and substructures, H-4arc plans increased V 5 and D mean of the heart, but reduced  Table 3 shows the results. T-8arc plans signi cantly reduced the V 5 , V 10 and D mean of lungs and D mean of liver compared to T-4arc and H-4arc plans, and no statistically signi cant differences were found in these indexes between the latter two plans. For V 20 , V 30 , and V 40 of lungs, there were no statistically signi cant differences between H-4arc and two Trilogy plans (p>0.05). H-4arc plans increased the skin doses (V 30 , V 40 , V 45 , V 50 , and D mean ) compared to two Trilogy plans.
T-4arc plans showed the largest number of MUs compared to T-8arc and H-4arc plans, and no statistically signi cant difference was found between the latter two plans. The average delivery time of T-4arc, T-8arc, and H-4arc were 6.05 ±0.57, 6.04 ±0.89, and 2.18 ±0.15 minutes, respectively. H-4arc plans showed the shortest delivery time and was statistically different from the two Trilogy plans. Table 3 shows the results.

Discussion
The design of treatment plan for synchronous bilateral breast cancer radiotherapy is di cult because of the large target range and the requirement of reducing the cardiopulmonary dose as much as possible.
Compared to conventional LINACs, Halcyon has shown comparable dosimetric distribution and excellent delivery e ciency in radiotherapy for tumors in multiple body parts. To the best of our knowledge, this is the rst report on the dosimetry of Halcyon for bilateral breast cancer.
Image guided must be took before each patient is treated on Halcyon. Four different setup elds delivered different MUs can be selected. In this study, the dosimetric differences of four setup-eld plans were rstly compared. Our study showed that CBCT-H plans, which delivered 10 MUs in each treatment fraction, increased V 5 and D mean of heart, LV, and lungs and V 5 of LAD and B-P. No statistically signi cant differences were observed in dosimetry among CBCT-L, MV-H and MV-L plans. In cases of delivering more MUs in each treatment fraction, the low-dose MV CBCT plans did not increase the doses to OARs compared to MV-L and MV-H plans. Therefore, we inferred that low-dose MV CBCT was the optimal setup-eld mode. Flores-Martinez et al. [6] compared four different setup-eld plans for unilateral breast cancer, in their opinion low-dose MV CBCT was the most suitable technique for patients treated on Halcyon. The result was the same as ours. Our and Flores-Martinez's study were all based on Halcyon 1.0. In Halcyon 2.0, KV CBCT was introduced, which displayed the lower dose and faster acquisition.
Second, the differences between whole and partial arc plans in the same LINAC were compared. The results showed that partial-arc plans on Trilogy showed more dosimetric advantage, especially in lowdose volumes of the heart, left ventricle, and lungs and the mean doses of the heart, lungs, and liver. This is because during designing a partial-arc plan, it is possible to arti cially choose the arc degree that irradiates less volume of OARs. Rotating the collimator angle and xing the jaw can further reduce the in uence of the leakage between the MLC on the dose of the OARs, which can minimize the dose of the OARs. Boman et al. [11] compared the dosimetric differences between the whole and partial arc VMAT plans of unilateral breast cancer including regional lymph node irradiation, and the results showed that partial-arc plans signi cantly reduced the dose of the ipsilateral lung and the V 5 of the heart, but increased the V 5 of contralateral breast. The result is similar to ours, but the cases of our study are bilateral breast cancer which does not involve the dose of the contralateral breast. Comparing the two plans of the Halcyon, the results were contrary to the Trilogy's, and the whole-arc plans showed better dosimetry. For PTV, in addition to the mean dose, the whole-arc plan was better than partial-arc plan in terms of the maximum dose, minimum dose, conformity and uniformity. For OARs, the partial-arc plan increased the dose to the heart, left ventricle, LAD, and lungs. The results showed that partial-arc plans have no advantage for Halcyon, which may be related to the jawless setting on Halcyon. According to the results above, when designing the treatment plan for bilateral breast cancer, we can choose a more suitable arc setting according to the corresponding LINAC.
Based on the results above, we mainly compared the dosimetric differences between Halcyon's whole-arc plan and Trilogy's two plans. All plans meet clinical requirements. For PTV, apart from the mean dose, there were no statistical differences in other dosimetric parameters between the two partial-arc plans on Halcyon and Trilogy. Compared with the T-8arc plan, H-4arc showed worse D mean , D 2 and V 107 , and better D 98 and CI of PTV. In short, the plans of the two machines are comparable in terms of target dose.
Darby et al. [15] found a linear relationship between the mean dose of the heart and the incidence of ischemic heart disease, and the incidence increased by 7.4% for 1Gy increase in the mean dose.
Therefore, the mean dose of the heart is often used as a reference for cardiac toxicity. However, the dose of cardiac substructure also needs to be considered in radiotherapy. Some studies believe that LAD and LV are important parts of the heart related to radiation-induced heart disease [16,17]. In this study, for lowdose irradiated volumes of the heart and LV, partial-arc plans on the Trilogy showed the lowest values, while for high-dose irradiated volumes, the whole-ac plans on the Halcyon showed the lowest values. For the mean dose of the heart, the partial-arc plans on the Trilogy showed the lowest value, and the wholearc plans on the Halcyon showed the highest value. This may be related to the additional radiation dose to the heart from each MV CBCT scan. For all dosimetrc parameters of LAD, there is no statistical difference between whole-arc plans on Halcyon and the two plans on Trilogy.
For lungs, the partial-arc plans on the Trilogy reduced the low-dose irradiated volume (V 5 , V 10 ) and the mean dose, but increased the high-dose irradiated volume (V 40  have shown that the in uence of contamination electrons in the FFF mode is greater, which further leads to an increase in surface dose [20,21]. Barsky et al. [5] retrospectively analyzed 34 breast cancer cases treated on the Halcyon, and the results showed that breast cancer cases were well tolerated on Halcyon, and the acute toxicity was comparable to the published reports of conventional LINACs. No grade 3 or higher acute toxicities occurred, and grade 2 dermatitis occurred in only 6 patients. For the liver, studies have shown that when the mean dose of the whole liver is <30-32 Gy, the incidence rate of radiation liver injury is <5% [22]. In this study, the mean dose of liver was less than 7Gy in all plans. We believed that all plans are safe in terms of liver protection.
Previous studies have demonstrated that Halcyon could reduce the treatment time signi cantly compared to conventional LINACs [3][4][5][6][7]. Our study proved the same results in radiotherapy for bilateral breast cancer. This may be related to the Halcyon's faster gantry and MLC speed and higher dose rate. The shortening of treatment time can reduce intra-fraction movement, improve patient comfort, and increase machine throughput.
In summary, our study could provide a reference for radiotherapy of bilateral breast cancer on Halcyon.
VMAT plans on Halcyon can meet clinical requirements in radiotherapy for bilateral breast cancer. The low-dose MV CBCT could be selected as an optimal setup mode. For Halcyon, the whole-arc plans are more excellent than partial-arc plans. Although Halcyon increase the mean dose of the heart and the dose of the skin compared to the conventional LINAC, it is comparable to the Trilogy in dose of the cardiac substructure and other OARs, and signi cantly reduces the delivery time.